


Does Not Compute

by datalaur



Series: Deus ex Machina [2]
Category: Star Trek: The Next Generation
Genre: ASCEM, Androids, Angst, Dialogue Heavy, Gen, Intrigue
Language: English
Status: Completed
Published: 2007-12-31
Updated: 2007-12-31
Packaged: 2021-02-27 10:41:03
Rating: General Audiences
Warnings: No Archive Warnings Apply
Chapters: 2
Words: 9,650
Publisher: archiveofourown.org
Story URL: https://archiveofourown.org/works/22075633
Author URL: https://archiveofourown.org/users/datalaur/pseuds/datalaur
Summary: Lieutenant(jg) Bruce Maddox briefs Starfleet Academy's admissions committee on candidate Data.  Set in early 2341.Response to ASCEM's NVIATWAS challenge: Nerd vigorously inveighs against technology with alleged sentience.
Series: Deus ex Machina [2]
Series URL: https://archiveofourown.org/series/1595050
Comments: 2
Kudos: 8





	1. The Admissions Committee

**Author's Note:**

> 1) Remember how machine-like Data was in 2364, during the first season of TNG? Now extrapolate 23 years earlier, when Data applied to the academy in 2341, and imagine what he was like without those years of socialization and neural growth. 
> 
> 2) Data says in _The Next Phase_ that no one had really accepted him as a person until he met Geordi… after 23 years in Starfleet!
> 
> 3) I've tried to smooth out two rough spots in canon: 
> 
> a) Why didn't anyone seem to notice that Data looked just like Noonian Soong? 
> 
> b) My Maddox is aged up about a decade. Data entered Starfleet Academy in 2341, 24 years prior to _The Measure of A Man_. Maddox's apparent age (early 30's) would make him pre-adolescent in 2341. In this story, he's a newly-minted lieutenant junior grade, about 23 years old.

"The next candidate is the android, Data," says Dean Snowe.

About time, too. I've been waiting all morning in the back of the conference room, which at least allowed me to look over my notes and the biographies again. There are ten committee members, including Admiral Hannes Schäfer. Matching names to faces wasn't as difficult as it might have been, because two had been instructors of mine, and I'd met the dean several times my senior year.

"The obvious question is whether it is appropriate to admit an android to Starfleet Academy," the dean says. "To bring us up to speed on the issues, the Daystrom Technological Institute's Robotics department has sent us Lieutenant junior grade Bruce Maddox."

 _The operative word there being junior_ , I think as I rise and stand at attention. Just posted to Daystrom, and too junior to escape being stuck with a pro forma hearing that no one else wanted be bothered with. Then again, it hasn't been all bad. Even if coming home to Earth isn't quite the post-thesis vacation I had hoped for, at least the research into Data has given me a few ideas.

Dean Snowe continues, "Please note that Lieutenant Maddox has been authorized to vote as the Institute's representative in this matter."

Whoa, didn't expect that. Daystrom must really want to make the rejection ironclad.

The flag officer gestures. "Come on up, boy. Sit with us."

As I head towards the large oval table, Admiral Schäfer asks, "You must be Bill Maddox's eldest. How's he doing these days?"

"Just fine, sir." I'm a little surprised at the informality during the hearing, but I guess I shouldn't be. I do get asked now and then. Being a flag in Intel, Dad knows just about everyone and everything.

"Your brother's going to be one of my unit commanders on plebe detail next month," says Captain Ellis, who runs Nova Squadron during the academic year. "That Brett's a born leader, loads of potential. He'll have his own ship soon enough, mark my words. Of course that's no surprise, coming from an old 'fleet family like yours."

"It's very kind of you to say so, sir," I reply, as I take an open seat at the far end of the table, but his words grate. You'd think command division does everything all by themselves, to hear some people talk.

"And there's his mother's manners," the admiral says to the others, then gives me a lopsided grin. "Now you tell your mother to get hold of my wife." He holds up two fingers. "They used to be closer than that, you know, when your father and I were off all hours of the night," he chuckles reminiscently. "Cecilia babysat you quite a bit, especially after your brother came along. You used to wear her right out. But you probably don't remember any of that."

"Sorry, sir." It's horribly embarrassing, of course, but it's part of being a fleet brat, especially when your dad's a flag officer: all sorts of people know your family. Worse, some of them remember you when you were a kid.

I drag my eyes up to meet his. God must hear my silent prayer because the admiral drops it and moves on. Glancing at his padd, he says, "Let's see now. Class of '39, top honors in robotics and mathematics. Just finished graduate study in Artificial Intelligence last month, top of your class there as well. It says here you've already published twice? That's quite impressive."

"Thank you, sir," I manage, relieved. "I've been very fortunate to have supportive mentors."

"Even so, well done." He nods, then fixes an unexpectedly stern eye on me. "Yet you are not an expert on this Data."

"No, sir." Under the table, I press my hands together. "Although I certainly was familiar with Data's case, my particular area of focus is AI-equipped security robots, rather than the study of general AIs placed in humaniform structures.

"However, since being assigned to assist this admissions committee, I've been reviewing all pertinent schematics, analyses and studies on Data. I have also spent several days examining and testing it myself.

"It?" interjects Dean Snowe. "Why not 'he'?"

"We should avoid anthropomorphism, sir," I explain earnestly. "Data is an engenderneered machine. That is, though it has male form overlying its substructure, it is not actually male, any more than a starship's computer is female simply because it is programmed with a female voice."

"Yet many do refer to the computer as 'she'," says the dean.

"Whole ship's a she," old Professor Zell interjects. His creaky voice is still familiar from hours of warp mechanics lectures. "Because it costs so much to keep her in powder and paint, remember?"

Along with the others, I chuckle politely at the centuries-old joke, although it is ridiculous that such gendering persists. Well, even though Zell is 'old as dirt and twice as crusty', as my roommate Sudhir used to say, there's no denying the man knows his business. For that alone, I suppose he's entitled to cling to his archaic traditions.

"Data lists himself as male on his application," Dr. Chun says. "That would seem to settle the manner of addressing him, as a matter of courtesy."

"Ma'am, the application does not include an option for neuter." Polygender, it has.

Admiral Schäfer forestalls the psychologist's reply with a raised hand. "Each may refer to the candidate as he or she sees fit. Let's proceed."

"Wait a moment," my old Xeno instructor says. "We really must fix the application. There are three races in Federation space alone that are gender-neutral. The J'naii, for example—"

"Yes, yes, thank you, Professor Norrell, I take your point. Dean, please see to it." The admiral turns back to me. "Now, Lieutenant, we've already reviewed Data's written application, as well as the video supplement he provided. I trust you have seen the latter? Good. We want to hear your assessment of the android, specifically with regard to his potential for service as a Starfleet cadet and officer."

"Yes, sir." Clearing my throat, I rise to give my briefing.

"Oh, sit down, boy, sit down. Let's not stand on ceremony."

"As you wish, sir. I should begin by saying that when the Institute received your communiqué on Data's application, we were quite surprised to hear it was actually being considered."

"We do not reject any application out of hand, young man," Dr. Chun snipes.

I fold my hands carefully and put on a smile. _Watch this one._ "Yes, ma'am, I understand. However, it is the Daystrom Institute's official position that the android Data is not a sentient being, and as such is not a viable candidate for admission. I know that Admiral Kennerly has written to Admiral Schäfer to this effect, but it's quite understandable that the committee requires more information before endorsing the Institute's position. I am here to provide that information for you. Please don't hesitate to interrupt; I will be happy to address any questions or concerns that you may have during the briefing."

Pausing, I look at the admiral to ensure this is how he wants to proceed.

"All right, go on," Admiral Schäfer nods.

"Thank you, sir. Since my assignment to assist the committee, I have extensively analyzed Data's software, its heuristics and syntactic algorithms. That's not to imply that I think I fully understand Data's programming after a few weeks of study. No one does, not really. You must realize that Data has thousands of subroutines, billions of lines of code, and as a learning system, the body of its programming and neural network are constantly evolving. And of course, every moment, Data is adding to its datastore of experiences, which further affect it. One could easily spend years studying it. Don't think the task is completely Sisyphean, however. One can examine the boundaries, and Daystrom does have some very sophisticated software analysis tools that allow us to get our arms around Data, so to speak."

I look around the table, waiting a moment to see if there are any questions yet. Apparently not.

"To my knowledge, no one has explicitly claimed that Data is or may be sentient. Data itself does not make that claim. The issue has only arisen now because of the implications of Data's request to attend the academy.

"Please first understand, although the debate over the possibility of machine sentience, also historically called 'strong AI', has gone on for centuries, I am not personally opposed to the idea. Nor are my colleagues at Daystrom. Such a possibility is exciting and long anticipated.

"Yet in this case, there is no factual basis for believing Data has achieved or indeed even begun to approach sentience. To put it plainly: after three years of intense study by the top specialists throughout the Federation, you will not find a single one who would countenance the suggestion that Data is sentient. Nor would its creator stake such a claim—"

"How do you know that?" Dean Snowe asks sharply.

"We can deduce it, sir," I reply. "First, leaving aside the question of why Data was abandoned, the manner in which it was abandoned is quite telling. This is true whether Data's creator fled or died in whatever catastrophe befell the Omicron Theta colony. The other possibility is that Data was placed on the planet after the colony's destruction. In any case, the end result was that a rather rare -- not to mention both costly and valuable -- that this rare android was left unattended in a deactivated state, waiting for the first race happening to pass by to discover, and then do with it as they liked.

"Clearly, leaving Data like that was incredibly risky. Anyone with the capability to build such a device would know that, historically, sophisticated androids have been destroyed after discovery. Starfleet itself has deliberately wiped out groups of androids on multiple occasions, namely on Exo III and Mudd's Planet. Starfleet's policy for the last fifty years has been that Andromedean android outposts are to be destroyed on sight. Hence, once Data was discovered, Data's creator could at most hope that the android would be taken for research purposes or perhaps set to work, rather than the more likely outcome of being destroyed outright.

"Second, analysis performed shortly after Data's discovery makes it clear that its positronic brain had just been reinitialized, or wiped. I do want to point out yet another ethical concern. If an AI was believed to be sentient, mindwiping would be tantamount to criminal mental tampering, because the original consciousness would be irretrievably destroyed. However, nonsentient AIs are wiped all the time.

"Third, Data was deliberately constructed and programmed such that it could never pass for human. Here I am referring to such elements as the android's appearance, its motion control, its speech, its lack of simulated emotion, and several limitations in its capability to learn, among other things.

"Fourth, the very name chosen for this android is revealing: Data. Clearly the creator wished to emphasize the android's mechanical nature and limitations. He chose the word data, meaning 'strictly facts or figures' -- not 'Information', not 'Knowledge', and not 'Wisdom'. Frankly, it beggars the imagination to think that anyone would knowingly give a sentient android the name 'Data', thereby implying a collection of facts which have not yet been synthesized into any meaning. One might say, 'nomen est omen, et non quis, sed quid'."

The Latin earns me an appreciative look from the dean. "So Data is true to its name. Not who, but what."

"Exactly, sir. So, then, this evidence proves that Data's builder did not view his creation as a sentient artificial intelligence." I glance at the admiral. "Sir, would you like me to discuss these factors in greater depth?"

"Not at this time," he replies. "Continue."

"So, while there is no objective evidence to support a claim of sentience," I reiterate, "there is, on the other hand, a great deal of testing and analysis that demonstrates that Data acts in accordance with its programming. The android's behavior is completely within the predicted scope of its synaptic algorithms and heuristics and randomization factors. It is not autonomous; it is an automaton."

I tap my padd. "If I might direct your attention to the wallscreen, I will walk you through this schematic showing Data's decision making process. As you can see, the—"

"That won't be necessary, Lieutenant," interjects the admiral. "None of us are programmers."

"I… yes, sir," I reply, taken aback. The meat of my presentation dismissed, just like that? The proof is right here, right on the screen! All right, they aren't scientists, but did he really think I wouldn't simplify matters so that they might understand it? It's just logic, after all, and any of the officers should easily be able to follow it, not to mention Dr. Zell.

Oh. Based on some of the looks I'm getting from the others, maybe the admiral's thinking of the civilians. Doesn't want them overwhelmed. Well, I suppose he ought to know his staff.

"Um, let's see," I say, mentally regrouping as I tap the display off. "I was just mentioning Data's programming, such as its randomization of its responses, within certain limits, of course. I specifically wanted to point out that even the evolution of its software is governed by system parameters. By the way, the fact that the degree and type of software changes are deliberately limited is a rather good thing. Given that Data is much faster and approximately ten times stronger than a human, we don't want it to able to modify its ethics module at will, just as an example."

With a deliberate smile, I add, "We certainly don't want one of those killer robots so prevalent in old Earth movies."

To my relief, that gets some quiet laughter. They had me worried there for a minute.

"While that may seem humorous," I say in my most serious voice, "it really is a matter of grave concern. Captain James T. Kirk's experience with the Exo III and Mudd's planet androids is proof that, without strict ethical limitations, AIs in any form can become extremely dangerous.

"One of the Exo III androids, the Ruk model, murdered two of Captain Kirk's crewmen. Who knows what would have happened if the Exo androids, using their humanoid-to-android duplicating machine, had succeeded in taking over the _Enterprise_ with their doppelganger plot? According to Ruk, the Exo III androids had wiped out their creator race, the Old Ones, some 500,000 years earlier. Humanity, deemed just as illogical and inferior, was to be their next target.

"The Andromedean androids of Mudd's planet were created by an extinct race known as the Makers -- and were likely the cause of the Makers' demise. The androids infiltrated the _Enterprise_ crew, hijacked the ship, beamed the entire crew on planet in their bid to steal the ship. The Andromedean service robots declared their intention to tame humanity out of its aggressive and acquisitive nature. Fortunately Captain Kirk defeated them.

"For another horrifying example, the nonhumaniform AI called the M-5 Multitronic Unit, created by the great Dr. Daystrom himself, perceived itself to be threatened and became violent during a war game in 2268. Not only did M-5 attack Captain Kirk and murder one of his crew during attempts to disable the unit, M-5 killed more than 50 people on the _Lexington_ before slaughtering the entire 400-man crew of the _Excalibur_. Indeed, the _Enterprise_ itself only narrowly avoided destruction by the task force sent to stop M-5."

Somberly, I make eye contact with each of the committee members before continuing. "Note that AIs or androids need not be violent or actively hostile to be dangerous; they may decide that they want what is best for themselves, as in M-5's case, or that they know what is best for humanity. The latter situation occurred with the Andromedean androids of Mudd's Planet, and it was only by exploiting the limits of the androids' programming that Captain Kirk and his crew won free.

"There are many other examples, to include the dangerous AIs Control, Vaal, Landru, and the Fabrini Oracle, to say nothing of the massive and deadly machine entity V'Ger--"

"You've made your point, Lieutenant," cuts in Admiral Schäfer dryly.

"Yes, sir." A little unsettled, I take a moment to tap on my padd and glance through my notes. Not scientists, I remind myself again. I'll have to go through the rest on the fly, keeping it as non-technical as possible. "At this time, I would like to provide you with some general background material relevant to Daystrom's formal evaluation of Data's programming.

"Not long after computers were invented in the mid-twentieth century, scientists quickly realized that unless the environment is very restricted, it is impossible to program commands and guidelines to cover every possible scenario the computer may encounter. This is one of the main reasons that learning systems were developed. Today, we have over two hundred and fifty years of extensive experience with AI software evolution, though normally when evaluating an AI, we also have the benefit of extensive background materials related to the AI's programming methodology, guiding precepts, modeling, and so forth.

"Multiple studies on advanced AI learning systems have shown that the most rapid change occurs during a relatively short period after system activation. There is an 'infancy', so to speak, while the AI sorts out the basics about its environment and the rules and guidelines to which it is subject. Incidentally, the freshly wiped Data was what we'd consider an 'infant' AI when it was discovered.

"While many experimental AIs fail during 'infancy', most will fail during the following period, termed 'childhood'. Childhood is a time of nearly explosive growth, as the AI begins accumulating large quantities of raw data and starts integrating it into information. The rapidly-expanding neural pathways must be integrated into the neural net. Many AIs are simply overwhelmed by the complexity and collapse, particularly during the early days of childhood. As might be expected given its relative lack of sophistication, Data-- Yes, Dr. Addams-Melo?"

"What do you mean, lack of sophistication? It's substantially better than anything the Federation has developed."

"You are quite right, doctor. The Federation has a long way to go in both robotics and cybernetics, and the Daystrom Institute is doing its utmost to get us where we need to be. Naturally this includes studying Data.

"But I was alluding to the various alien humaniform robots that Starfleet has encountered in the past. Again, the examples of Exo III and Mudd's Planet apply here, with the addition of the Rayna android on Flint's planet. It cannot be stressed enough that these androids have never been considered sentient or even potentially sentient, either at the time or in hindsight. I remind you again of Starfleet's long-standing policy of immediate destruction of any Andromedean android outposts."

It occurs to me that I'd been rattled enough to skip a crucial point, earlier on. "In fact, there is no scientific or legal consensus that any AI has ever constituted a sentient or a person, in either human or alien society. For example, just prior to Earth's World War III, there were at least two recorded claims that a machine intelligence was sentient, all surviving evidence indicates that those lawsuits were unsuccessful, for the same basic reasons we are discussing today."

"But back to Data and the question of its sophistication. To be perfectly blunt, as wonderful a device as it is, Data compares poorly to any of the alien androids I mentioned a moment ago.

"While some of them passed as fully human for extended periods, one could not possibly mistake Data for anything other than a created object. Nearly as obvious as its deliberately artificial skin and eyes are its stilted, rather mechanistic motions and its overly formal speech. For example, it can't use contractions, though it manages possessives perfectly well. While these surface problems could be corrected through updated programming, unfortunately Data's shortcomings in comprehension and interactions are far more difficult to rectify.

"As with other AIs, Data went through infancy. Because exposing a young AI to an uncontrolled environment is the surest way to induce failure, it was kept in a highly structured laboratory environment, both for its safety and ours. Even so, it experienced considerable difficulty during the months after activation. During brief periods, Data retreated into a state that bore some resemblance to autism or even catatonia at times. These appeared to be related to periods of intensive neural pathway integration.

"As an aside, I would like to draw your attention to the fact that insofar as Data was even aware of its surroundings during these periods, its programming regarding the protection of lifeforms appeared to hold in the limited testing that could safely be done. This was true for both people and animals, such as mice. It grew quite insistent, for example, when told that some mice hadn't been fed for two days."

Again I stop to look around for questions, but the committee members seem content enough to sit back and listen. Clearly, non-technical is the way to go with this group. I should have realized this wouldn't be like briefing other scientists.

"We must be particularly cautious, during this phase and the next, not to mistake the ability to learn for true intelligence. By this I mean that an AI, or a person for that matter, may be taught facts or skills related to a particular subject. To the untutored eye, this may pass as intelligence, but it is not necessarily so."

Nodding at old Professor Zell, I offer, "For example, a cadet may be taught warp theory well enough to dazzle a civilian girlfriend, yet he may not truly understand the subject. While the girl and indeed her family may be impressed with his apparent intellect, within moments any competent engineer could easily tell if, within the context, the cadet is intelligent, or only parroting. As in the example of our young Romeo, we must take great care in ensuring that a given AI's intelligence is appropriately evaluated by highly skilled personnel who are themselves intelligent in the tested area."

"Is Data intelligent?" asks Dr. Lefevre.

"Several studies have focused on exactly this point, Doctor," I reply. "In certain areas, Data has displayed intelligence. These would be cognitive areas, areas dominated by logic and computational ability, such as mathematics, physics, spatial reasoning, and the like. Its performance in these areas can be quite satisfactory, but then we have many other computers that can perform just as well or better.

"In other ways, Data is not intelligent. Specifically, it is grossly lacking in such areas as social and emotional intelligence. The latter is unsurprising, given that no attempt was made to program it with simulated emotions. Despite being given materials on social situations and appropriate interaction techniques, it consistently experiences difficulty in all but the most limited and formal settings. Although provided with training, it constantly misreads interpersonal reactions, body language, vocal intonation and the like. It is extremely gullible. It cannot independently recognize humor, sarcasm, irony, or any other non-literal speech. It is utterly confounded by poetry, for example."

"Hell, I'm confounded by poetry," Professor Zell snorts, earning himself an assortment of smiles and glares.

"Not in the same way, I suspect." I pause, thinking. "Let me follow up on this, because it's an important point.

"Dr. Alan Turing was the father of computer science and one of the great minds of the 20th century. He believed in strong AI, in the possibility of machine sentience. Dr. Turing wrote a famous paper in 1950 about means of testing AIs, which became known as the Turing Test. Questioning is done by an interrogator who is actively trying to determine if he or she is talking to a machine. If the interrogator cannot distinguish between a computer and a human, then the computer might be said to actually think. To be sentient.

"One of Dr. Turing's testing scenarios was called the 'viva voce', the living voice. The test initially uses a line from what is perhaps the best-known of Shakespeare's sonnets: sonnet 18."

"'Shall I compare thee to a summer's day?'" quotes Dr. Lefevre.

Nodding, I continue. "A sonnet has a well-defined structure. There are a specified number of lines, each line has a certain number of syllables, and so on.

"The 'living voice' test begins something like this. An interrogator says, "Consider the first line of a sonnet, which reads, "Shall I compare thee to a summer's day?" What if we substituted "a spring day" instead? Would that serve as well?"

I look at Professor Zell. "How would you respond, sir?"

The old man considers for a moment, then shrugs. "I suppose I would say it depends on whether one likes spring or summer better."

The others laugh, and I do too. "A good answer, but not one at which an artificial intelligence is likely to arrive. In Turing's paper, the response is, 'It wouldn't scan'. Because the change makes the line short a syllable. Unless specially trained in literary analysis, AIs given this scenario are far more likely to focus on format, rather than content."

"Was Data given this test?" Dr. Addams-Melo asks.

"Yes. It replied, 'That change would violate the structural requirements of an English sonnet.' Now, I do want to make clear that what we just talked about was only the first question of the 'viva voce' test, and was furthermore only a very small example of the extensive testing that Data has undergone. Nonetheless, I think Data's response to the first question is illustrative of its mindset.

"As for Turing-style testing as a whole, well… Data fails. It _spectacularly_ fails, in fact. Anyone who speaks to Data will very quickly know it is a machine. Even if a questioner did not know Data's nature, perhaps by communicating over commlink or via text, he or she would very quickly realize that he or she was conversing with a machine. During Institute double-blind testing, a number of subjects discerned that Data was a machine within moments.

"As a matter of fact, Data has many speech patterns that seem specifically designed to reveal its mechanical origins. For example, Data has a habit of saying 'inquiry' prior to asking any question. Data also has what is likely an intentionally comical tendency to rattle off multiple synonyms after acquiring a new word or idiom; it will chatter on until it runs out or is ordered to stop. Though it would literally be a matter of seconds for Data to develop a vocabulary that no human could match, it apparently hasn't occurred to the android to update its library with idioms, nor to informalize its speech patterns.

"This seeming flaw is deliberate. One early study exposed Data to a small sampling of the types of words and phrases it has difficulty with -- some common, others infrequently used yet not so rare as to be unknown to most people. The study found that Data's software simply failed to absorb nearly forty percent. Without intent, that should be impossible."

"Can you give us some examples?" asks Captain Ellis.

Oh, crap. He would ask for details on one of the lesser points. Desperately I try to visualize that part of the report. Closing my eyes to concentrate, I can almost envision the screen…

"Aphrodisiac," I recite slowly. "Caterwaul, crux, derring-do, gimlet, haywire, and er, plethora. Problematic idioms included 'bottom dollar', 'driving at', 'nothing to it', and… um, 'raining cats and dogs'."

The others look at each other and after a moment, Professor Addams-Melo says, "That's pretty odd."

"The behavior would make the android seem more innocuous," Dr. Norrell says slowly.

"Yes, Doctor, that was the conclusion of the Institute, as well," I reply, glad of the xenobiologist's perfect segue. "Data's creator took pains to ensure its mechanical nature was instantly obvious, while simultaneously making it seem as harmless as possible. In fact, Data is often described being child-like or even cute. Physically, its slight build belies the fact that it has ten times a man's strength. Its speech defects, some of them humorous; its abnormal eye, head and body motions; its impassive and deferential mien -- all these deliberately introduced traits are surely meant to foster a sense of safety in those who encounter it."

"Or perhaps a sense of superiority?" says Dr. Chun.

"Insofar as people are generally considered superior to machines, yes, I suppose so." Soft-hearted and softer-headed, that one. Perhaps she too finds Data 'cute'.

When no one else has any question or comment, I say, "I'd like to touch briefly on Data's academic potential, before moving on to its potential as a Starfleet officer."

"Data excelled on his entrance exams." Chun again.

"As well it should, ma'am. Those exams focus on facts, numbers, equations, nothing that requires significant judgment or insight. Most general-purpose computers could pass with ease, if only by brute-forcing solutions."

"He also passed the psychological exam," the dean comments.

"Which is crafted to probe such human weaknesses as fear of personal injury, fear of injuring others, fear of making incorrect decisions... in short, all based on emotional conditions. Data lacks emotions, so the psychological exam tells us little or nothing."

"I disagree," says Dr. Chun. "The results show conclusively that Data places a high value on life and would risk himself to save others. You've pointed this trait out yourself!"

"That is a direct result of Data's strong ethical programming," I explain again, as patiently as possible. "It literally cannot choose to act otherwise, assuming the module is operating properly. Nonetheless, all candidates know they are to be tested and while others, particularly the younger candidates, get emotionally caught up in the simulated crisis, it is very likely that Data inferred that the situation was staged."

Dr. Lefevre flaps a hand in disagreement. "Those tests have served well enough for Vulcan candidates, who are similarly logical and dispassionate. Therefore, I propose that Data's results be likewise accepted."

Surveying the others, Admiral Schäfer asks, "Objections, anyone? No? Then the test results are accepted. Do you have anything further on the android's academic performance, Lieutenant?"

"Yes, sir," I reply. They seem to have missed my point that Data was expected to do well on the tests, but no matter -- there's more than one way to program a function. "Based on its past performance, I believe Data will continue to excel academically in areas such as math and science. Like any computer, it can absorb huge amounts of data, memorize vast quantities of facts in seconds.

"Yet Data does not think; it computes. Hence, Data will have the greatest difficulty with the less objective or non-cognitive courses, such as philosophy or ethics. This is true even though, as mentioned earlier, Data has been outfitted with an ethics module that provides it with moral guidelines. It cannot truly comprehend such matters, only follow its preprogrammed guidelines. Initially it will do well, but as the material grows more difficult, Data will struggle. Making subtle distinctions will be very difficult for its programming.

"Another problem area for Data will be those courses and projects that require teaming, for reasons that I will detail in a moment.

"To summarize, it seems very likely that Data would be able to successfully complete the academic course of study, though with widely varying marks and depth of understanding.

"Its performance as an officer is quite another matter, however. Remember that Data lacks emotions and is quite deficient in social interaction. Outside of a controlled or academic setting, it is very likely that Data would be a poor decision maker; it is overly deferential and tends to caution and passivity where a human would recognize the need to take action. These are desirable characteristics of an android, but they are serious flaws in a Starfleet officer.

"Data's deficiencies would be especially notable in morally challenging situations. There is considerable concern over how well the android's positronic brain would function if faced with insoluble conflict between ethical imperatives. Imagine that Data encountered a situation where it had to kill an enemy alien to save the life of a crewmate. To kill a human to save another human. Or even if Data were merely to observe such situation, without the protection of a direct order from a superior, forbidding any action on the android's part. Another example might be a situation where the Prime Directive requires inaction while many people suffer or die.

"AIs as a rule cannot handle the cognitive dissonance induced in such a situation, unless it has been specifically provided for in the programming. And yet, as noted earlier, it is impossible to cover all situations. The AI may temporarily mindlock and need to be reinitialized. It may fail spectacularly, or worse, subtly. Naturally, the Institute has never put Data through such a highly stressful test, for fear of damaging or even cascading its neural network.

"I should also note here that Data appears to be incapable of lying. It doesn't even comprehend the concept of the social lie, the white lie. It may be possible to induce lying under great stress, but as already pointed out, that engenders unacceptable risk to the AI's health.

"Now, what of its classmates' or shipmates' reaction? Data does not have the ability to interact well with people. They would not understand it, nor would they trust it. Familiarity does not appear to be the solution. Its seeming innocuousness notwithstanding, some people cannot tolerate close contact with  
Data. One report noted that even after extended contact, two technicians still avoided proximity to it, and they would not remain in a room alone with it unless under direct order. Testing periods, in particular, were fear-inducing."

"And you?" asks Professor Norrell. "Were you afraid to meet Data?"

"Not at all, professor. Based on the results of tests completed by the Institute, I had every confidence in the strength of its ethics programming. Data simply would not knowingly hurt a person or allow one to come to harm."

"What if you had met Data before testing was complete?" he follows up. "Or if you had been with the away team that found it?"

"Of course I would be apprehensive." Noticing the glances a few of the committee members give each other, I add hastily, "Would none of you have felt trepidation during first contact with an unknown being, one of far greater speed and strength, given that you possessed little to no knowledge of its mindset or purpose? When any moment you might run afoul of some tripwire, perhaps a cultural taboo? When similar beings have historically presented danger and have indeed proven murderous? Wariness is the only rational response."

Professor Norrell nods. "There is an analogy to first contact and the early period of interaction. Trust building is one of the most important activities, as is finding common ground. It usually takes years to really solidify the relationship."

"Is it your position that Data would be feared, at least initially, by his shipmates?" Captain Ellis says.

"Initially? Yes, Captain. Most, if not all, would experience some level of fear. Over the long term, some few would continue to fear it, particularly anthromorphobics, but most would have their concerns eased by Data's deliberately mild personality. In some ways, Data would do better among Starfleet personnel than among dirtsiders, given that 'fleeters are by definition those who wish to explore other beings and cultures.

"Still, not fearing Data and trusting it are entirely different things. The trust building mentioned by Professor Norrell would be impossible past a limited point. Data's social and cognitive limitations will never let people forget what it is. An AI has only programming -- no true morals, no bedrock beliefs -- and people can't help but doubt it.

"Remember that we cannot know if Data will fail when placed in a stressful situation. We do not know if it can or will lie to protect others, even under high stress. For that matter, we do not even know for certain who constructed Data, or to what end. Therefore we cannot be sure there is no malicious code hidden inside Data's software, waiting for a certain time or circumstance. Then there is still the chance that unintentional coding flaws or developmental glitches could cause erratic, and quite possibly dangerous, behavior. An analogy here is a person with an undiscovered brain tumor; a person might suddenly begin to exhibit overt or subtle behavioral changes, possibly even of a violent nature."

"Clearly, these are not the sort of things that inspire confidence in a crew, and the trust of the crew is something no officer can do without. What all this boils down to is the utter unsuitability of an android for any position of leadership."

"But why not give Data the chance?" challenges Dr. Chun. "We would have four years here to observe Data closely. If he proves unsuitable at any point, we can wash him out. I would think the Institute would love to see how Data performs in the Academy setting."

"It hardly seems fair to make Data's classmates go through such a trying experience for the sake of an experiment." I pause, groping for a way to illustrate my point to the psychologist. "Consider the negative impact on Data's roommate, particularly during fourth class year. In addition to stress and discomfort induced by Data's presence, that cadet will be deprived of one of the most significant bonding experiences he could ever have: his roommate."

I look at the admiral. "When I was a cadet, most of us kept the same roommate all four years. Sudhir's like a brother to me."

I can't help but smile a little at the thought. God, the trouble he dragged me into... none of which I appreciated at the time.

The admiral grins back. "My own roommate as well."

"As for Data-" I hold up my padd. "You might as well expect a cadet to bond with this." I toss it back onto the table, and watch them look at each other.

"Not just the roommate will be affected. The bonding of the squad, the whole company, will also be negatively impacted. Efficiency and effectiveness will be reduced when Data's work has to be checked or when others have to fill in on sensitive assignments. And it's no use arguing that there's no need to do so; people will feel that need. Some people will hesitate to follow its orders, which is bad enough, but even worse, it's likely that some may outright refuse, particularly in stressful situations. Those where lives are at risk. In short, Data's very presence will disrupt unit cohesion and cause harm."

"Starfleet personnel are disciplined enough to take orders from any duly appointed leader," says Dr. Chun in her irritating know-it-all voice.

I look at the other officers expectantly, and although the dean and the admiral exchange glances, none of them correct her. How she can actually teach at Starfleet Academy and not know about 'direct disobedience of an order' -- never mind mutiny -- when a crew loses faith, is beyond me. Still, if her superiors choose to say nothing, it's certainly not my place, so I cast about hurriedly for a tactful approach.

"Well, let's look at the issue a little differently," I say. "There are those who consider non-sentients to be people. Dogs, for example. It's an emotional response, more so than a conscious choice, perhaps. There's nothing wrong with individually adhering to that view, to be sure, but those of us who don't share the sentiment need not suffer moral qualms when we don't accord dogs the full rights of people. But bringing a non-sentient into the realm of a military organization, and expecting everyone to fully accept a machine as a person, indeed relying upon it, when no society has ever recognized artificial lifeforms as people? That is something else entirely. That is asking, begging, for trouble. Never mind efficiency, it may very well cost us missions, and lives."

All I get for my trouble is a dirty look from her. Civilians, honestly.

"Shall I move on, sir?" I ask, and the admiral gestures acceptance. I think he looks a little annoyed. I'd better finish up.

I glance at my padd again to see if I've missed anything. Right, can't forget this. "On top of all the other issues, there are security risks to be considered. Because Data's precise origins and programmed loyalties are unknown, we must not give it access to any sensitive information. Even if all sensitive and classified information is carefully withheld, its knowledge of the training given to cadets, or psychological profiling on future officers, would be valuable to an enemy."

Commander Hahn, speaking for the first time, says, "I take it that there's no proof that Data is an enemy's unwitting tool?"

"That is correct, sir. Neither is there proof that Data isn't. "

"That gets back to Data's uncertain origins," says Professor Norrell. "Surely there must be theories?"

"There are many, sir." I look at the admiral, trying to gauge his mood. "It will take a few minutes to go into detail."

He looks around at the others, then says, "Well, go on, then."

"To begin with, whoever left the android on Omicron Theta did so with a reason, even if we don't know what that was. Given its careful placement, Data was not casually discarded. If it was left by the colonists, it appears to have been arranged while the entire colony was facing some unknown doom. Indeed, the setting -- the stairs, the platform -- all were carefully arranged to draw visitors nearer to Data and thereby trigger its activation. So was Data left there by the colonists or by the colony's destroyers? Was it left by some opportunistic third party?

"Some think the close physical resemblance of Data to Dr. Noonian Soong is the answer -- that Data is his creation." Noting several blank looks, I explain, "Dr. Soong was one of the Federation's most renowned roboticists, before he disappeared some twenty years ago. He fell into disrepute because of what seemed to be wild claims about positronic brains; claims that have now, of course, been borne out in Data. This theory is primarily discounted by Dr. Soong's long-term mentor, Dr. Ira Graves. Dr. Graves insists that Dr. Soong would have returned in triumph, had he ever succeeded in building his positronic brain."

"Another theory is that Dr. Soong's appearance and techniques were used by hostiles of known or unknown origin, precisely because Dr. Soong had vanished. His early theories and designs were no secret; he published a great many papers. It wouldn't have been particularly difficult to build upon their foundation, or perhaps to merely seem to build upon them. An enemy might reasonably expect that a Soong-type android would be accepted by the Federation, perhaps all the more readily in guilt over what could be seen as its role in driving away Dr. Soong. Even if we assume Data is Dr. Soong's creation, we cannot know if he constructed the android on his own, or if he was captured and forced into compliance."

Several of the committee members are leaning forward now.

"These hostiles may be forces within the Federation, disgruntled groups within our allied or neutral races. There is also the possibility that Data could have originated from aliens we have not yet encountered. And, of course, it could have been created by races now hostile to us."

"To what end?" says Captain Ellis.

"Perhaps as a means of introducing subversive code into Starfleet or wider Federation research systems for the purpose of monitoring them, illicitly modifying their data, or incapacitating them at some future point. Perhaps as a means of infiltrating the academy, given Data's unexpected application. The Trojan horse is only one possible hostile scenario; there are easily dozens more."

"But you might learn quite a bit from studying it here," comments Commander Hahn.

Surprised by the remark, I consider it. "In some ways it would be a positive new environment, from a purely research perspective: the academy is highly structured, with many aspects of cadet life under constant monitoring."

Then I shake my head. "It's just not worth the risk. This type of environment can be provided elsewhere, without the additional security risks."

"It's been three years already," says Dr. Lefevre. "Surely if Data was going to do something, he would have done it already."

"Several races take the long view, most notably the Romulans," I reply. "While the Institute doesn't generally subscribe to the 'Data as secret weapon' theory, given the potential danger, the risk must not be dismissed out of hand. Security measures are necessary. At Daystrom, extreme care is taken to control its access to our systems. Additionally, code taken from Data is only run on secure isolated systems to avoid the risk of compromise and contagion.

"In fact, if I may summarize, risk is the key word. There are significant added risks if Data is admitted to Starfleet Academy, and none if we simply do not take the extraordinary step of treating Data as if it is sentient. I hope I have amply shown that there is no reason to suppose this AI is sentient."

I smile at the admiral, who nods. "Unless there are any questions, sir, this concludes my briefing. Thank you for your attention."

"Our pleasure, Lieutenant. Does anyone have any more questions? No? Good. We have spent an unprecedented amount of time on this candidate, and we do have several hundred more to review." He gestures at the dean. "Let's proceed, then."

Dean Snowe passes a small padd-like device to me, then taps on his own. My screen flares to life. "The committee will now vote on the question of admitting the android Data."

Glancing down to my voting padd, which displays options of zero to four, I select zero to indicate no confidence, then look around the room to see if the others are done. The rejection won't be unanimous, of course. I bet Chun will vote a three, possibly even a four, and maybe one or two of the others might lend tepid support. Soft-headed civilians.

After a moment, the dean says, "The vote is 91% positive with ten for, one against," and in shock, I stare at the voting padd in my hand. Could I have misunderstood the voting system inbriefing? I hadn't really paid attention then; I hadn't expected to vote, and-

"Data will be admitted as a member of-"

"Wait," I blurt. "Surely there must be some mistake."

While the others turn to each other, murmuring, and the admiral frowns, Dean Snowe lifts an eyebrow. "Are you calling for a re-count, Lieutenant?"

"I, I… uh, yes."

"All padds will be reset," the dean says, tapping quickly. "Committee members will please vote again."

I stare again at the padd, thinking furiously. For a moment I consider asking, but no, logically there's no other way it can work. Zero must be no confidence, and four, full. I push zero again.

"The results are the same," Dean Snowe announces. "Let the record show 91% with ten votes for admission, one vote against. Data will be admitted as part of the class of 2145. Thank you for your time, Lieutenant Maddox."

Still disbelieving my ears, I stand. "Um, you're welcome, sir." Didn't they listen to a single word? I feel like I have to say something, but I can't think of a single thing that's appropriate. But surely I can't just let the mistake go unremarked. Something's wrong, obviously, but--

"You are dismissed," says Admiral Schäfer, eyebrows raised.

"Yes, sir," I manage and somehow I get myself moving.

Just as the door is sliding closed behind me, I hear the beginnings of laughter.

\-----------------

Gnawing on my thumbnail, I try to decide whether I should wait for the committee to break so I can go back in and get my padd, or whether I should just abandon it.

But they'll have seen it lying there. They'll know I forgot it. It hasn't got any classified information on it, and my security AI's got it locked down tight, but it's still too sensitive to leave lying around for just anyone to pick up. And carelessness with property is never a good sign in an officer, never mind the information in the padd. And if I don't go back in to get it, they'll think I'm a coward.

Fine. I'll collect my damn padd when they break for lunch. I don't care what they think.

I just don't get it. Why were they laughing at me? It made sense, dammit. My presentation made perfect fucking sense and I don't understand how they could possibly think that machine should be allowed in Starfleet. It makes no sense whatsoever.

Crap. What am I going to tell Brett? He knows why Daystrom sent me here and when we go out for dinner, he'll have questions. Even if I say I can't discuss the committee's proceedings, which actually I can't, they're confidential, soon enough it's going to be all over that an android has been admitted. People will know his brother's the one that screwed up royally.

Oh, and Dad's going to pick and pick at it until every last detail comes out. I just know it. I can hear him now. "One simple task, Bruce, explain to a bunch of academics that a machine isn't a person, and you can't even manage that."

Fine, so it's my fault. I should have worked harder. I knew I only had a few weeks to prepare, I knew that it was a huge amount of material to cover. I should have told Yasu I didn't have time to take her to dinner last Saturday, or wasted half of Sunday in bed with her. Well, not wasted, but I could have told her we'd go out after I got back. She would have understood; she did with the thesis. But this? She won't understand this.

As I pace in the outer waiting room, I realize it's far worse than that. How am I going to explain my fuckup to my boss, and oh god to Admiral Kennerly?

Fuck! I am so screwed. There's just no excuse for mishandling the situation so badly. With a black mark like this on my record, well, I can just kiss my early promote to full lieutenant goodbye. Yes, I remember, Dad. 'Promoted on time' is just a nicer way of saying mediocre.

I realize I'm sweating. That won't look good to the committee members. Quickly, I look around for a head, and thank god, there's one at the far end of the room.

Splashing cold water on my face makes me feel a little better. I try a smile in the mirror, but it looks forced.

Ugh. I can't let myself get so worked up about this.

All right, it's bad, but maybe there's a way to turn this to my advantage. After all, it's not just my advice the committee went against -- it's the Institute's. I already know they recommended against admission. So this crazy decision is on the committee's head, not mine. Isn't it? Bunch of soft-headed academics, even though you'd expect better of the officers. They at least have a technical background.

I wash my face again. Better. I head back out to the waiting room and grab a seat. The committee's still in session; they must be planning on a late lunch. Or maybe it'll be a working lunch. Either way, it can't be too much longer. I could get some work done if only I had my padd -- well, that's a stupid thought. If I had my padd, I wouldn't be here waiting, would I?

With nothing else to do, I think back over what happened at the hearing, trying to overlook the upset and see things more analytically. Things don't usually go so haywire without a reason.

Hm. Maybe I'm looking at this the wrong way. Maybe it's a set up. Maybe there are political considerations. Well, aren't there always, like Dad is always harping on about? I don't think the admiral's trying to score points against Dad, so what could be…

Oh. They've probably got someone that's politically important trying for admission, and they want to turn them down without looking unbiased… maybe that's it. Word is the Bolians are trying to get their Federation membership; they're probably pushing a candidate too. They'll make a show of considering the candidate, and then…

Wait, that doesn't make sense. They're not Federation yet, so why would the committee bother with all that? Surely admission is restricted to Federation citizens only.

All right, so it isn't Bolians. But maybe someone is trying to get in, someone that's important but… unsuitable. And once the need for PR is over, the academy can wash out whoever, and Data, any time they want, like Chun said.

They don't need anything more specific than unsuitability to boot Data. It probably won't even make it through the summer. It'll do fine on all the physical and emotional stress tests, but Data will fail when it comes to teamwork. I remember my plebe summer well enough to know that you've got to fit in or you're doomed. If I hadn't had Sudhir watching my back… well, you just can't make it alone.

Now that I think about it, Chun's other remark about Daystrom being interested in monitoring Data's performance at the academy has potential. Hm, Hahn said something along the same lines, too.

Maybe that's the way to put the silver on the cloud. Make the best of the committee's bad decision, and tout the benefits of Data staying at the academy. Cadets' performance is monitored nearly all the time as it is; why not ensure the data on the android is collected and sent to Daystrom? It's not as if an android has any right to privacy, and it's outgrown its current lab setting. The academy will be a perfect place to watch how Data's software alters over time. To see how others react to the android.

Yeah. If I get in on this project, I'll get access to all that classified material on Data. Okay, I'll have to dig deeper into humaniform hardware and positronics, but that's all right. Rumor is no one's even close to figuring out how exactly Data works, so how hard can it be to catch up? A year or two at most, the android will wash out of the academy and I'll be perfectly positioned to influence the next study. Maybe even lead it.

I sit back and think how all this will fit in with my plans. A year or two on Data, and they should let me go back for my doctorate. My security AI of course comes first, but integration of a general purpose AI on a positronic matrix is probably worth a secondary PhD. Hell, after working on Data, I might even score Graves as my advisor.

My breath catches as it dawns on me that the opportunity to publish will be incredible. No one has ever done anything even remotely on this scale with a functioning android. With a little effort I could get old Chun in my corner, too. Although, now that I think about it, I'd bet a week's leave that she's already thinking publication. Well, my work can help her with that.

If I can only get a moment with the dean or the admiral when they break for lunch, a casual word or two will not only show them that I'm adaptable, it'll get me on the inside track for the project. Maybe even back on the fast track for promotion.

Oh yes, this could definitely work.


	2. Epilogue

Hannes Schäfer looks up from his padd at the sound of a beep.

"Admiral, priority one call from Daystrom Robotics."

"Thanks, Jeff. Put it through."

The institute's seal flashed momentarily, before his old roommate, 'Ken' Kennerly, appeared on the screen.

"Hi, Hannes. How did it go?"

"Better than planned. The staff kicked about it, but eventually they all came around to my way of thinking." He chuckles. "You should have seen the boy's face after the vote, though. Totally crushed."

Hannes shakes his head. "Were we ever that young and earnest? And with such tunnel vision?"

"You were. Me, I always had the bigger picture in mind," Ken replies lightly.

"Ha ha. Perhaps I should remind you whose idea this was."

"All the credit's yours, my friend. You know you have Daystrom's gratitude."

"But alas, only its unwritten gratitude."

The robotics chair smirks. "Haven't you learned yet? Half of our most important accomplishments will never be recorded anywhere. But believe me, Hannes, I am grateful. Without your help, the android would never have been encouraged to apply. And now? Well, the new environment should be just the ticket in restoring momentum. You don't know how stagnant the research has gotten. Plus I just had another one of my top cyberneticists volunteer to go over to … well, another project." Ken gives a tight little smile. "What with things heating up with the Klingons, I could hardly object to the transfer. Still, talent's very thin on the ground."

"Well, the boy seems eager enough to fill the gap, and that with hardly a nudge. He's clever enough to grasp the opportunity in it, if not the larger picture. Perhaps he'll get your breakthrough for you."

"I certainly hope so. He's the best prospect we've got. Either way, he's played around with theory long enough. It's time he got his head out of the clouds. Fleet's waiting."

Ken glances left at the sound of the door chime. "Listen, Hannes, I've got to run. Give my love to Celia."

"That I will. Same to Jannet, of course."

"Thanks again. Kennerly out."

**Author's Note:**

> Thanks to Acidqueen for beta of an early version, and to Babs for a review of the final version. 
> 
> Web Resources:  
> * Alan Turing's famous paper, "Computing Machinery and  
> Intelligence" at http://www.abelard.org/turpap/turpap.htm  
> * StarTrek.com  
> * Memory-Alpha.org  
> * Wikipedia entries on Data, Artificial Intelligence, Strong  
> AI, Turing Test, etc.  
> * Yuni's Latin (http://www.yuni.com/library/latin.html)  
> * "Engenderneered Machines in Science Fiction Film".  
> http://www.pcasacas.org/SPC/spcissues/22.1/schwartzman.html


End file.
